Incorporating scholar's background knowledge into recommender system for digital libraries

Abstract

In recent years, recommender systems have received increasing attention in digital libraries since they assist scholars to find the most appropriate articles. However, a major problem of such systems is that they don't subsume user background knowledge into the recommendation process and scholars have to manually sift irrelevant articles obtained in response of queries. Therefore, a great challenging task is how to include scholar's knowledge into personalization process and filter out articles accordingly. To address this problem, a novel cascade recommender framework which incorporates scholar's background knowledge using ontological concepts into the user profiles is proposed. The framework exploits standard ODP structure as ontology modeling as well as lexicographic database (WordNet) for concept disambiguation. The primary experiment over CiteSeerX digital library indicates an increase in user satisfaction.